Most of the application scenarios of self-checking are optimization iterations, or revision iterations, which are based on the optimization of existing modules.
The last article shared with you the important influence of product design, and there are many ways to improve product design capabilities, and they are often used in combination, one of which is “self-check”.
1. Are you emotional or rational?
This is the question of the previous article, and it is also the beginning of this article. I hope that through these two articles, you can find your own answer. Of course, at the end of the article, I will explain several common data with you.
To make a product that we want is the long-cherished wish of most midwife managers. In actual work, we often use our own good or bad criteria to judge a demand. Over time, it becomes that we are the goal of this product. Users, others are “not target users”.
Product managers are not CEOs. Frankly speaking, we are farther and farther away from the CEO. This position is becoming more and more professional. In terms of vision, how many product managers are making the products they want?
- Product managers who like to travel may be doing financial products.
- Product managers who like movies may be doing information products.
- Product managers who like to make friends may be making some instrumental products.
In reality, the position and our vision are more mismatched. If we still carry out our work with this mentality, the risk factor is actually very high. It doesn’t mean that we need to match our vision and position, but we need to adjust our mentality from perceptual to rational.
Of course, if your position and your vision can match exactly, naturally there is no need to change, and even change from rational to perceptual.
That is, to make a product that the market needs, we need to be very cautious, not only at the demand level, but also in our design methods. How to use the technology of our product design to achieve the result of the data level when the quality of the demand cannot be judged.
- The methodology of self-inspection is used to find answers to the following questions:
- Is the new design really better than the old design?
- Do we have clear expectations for the new design?
- Has the new design met our expectations?
- and many more.
You may have noticed that most of the application scenarios of self-checking are optimization iterations, or revision iterations, which are based on the optimization of existing modules.
The reason is that the 1.0 stage of each module is in the verification, and the subsequent optimization is the self-check.
Regarding the verification method, I will have the opportunity to share it with you in the future.
2. Self-inspection methodology
This is a practical methodology that can be used directly. You can try to use it yourself or recommend it to your team. You can use it directly or adjust it before using it.
The self-inspection is divided into three stages: data observation period, data verification period, and analysis period. Relying on our understanding of data, we can visually analyze the improvement and decline of new and old products. Of course, this requires that our products have already done data burying points, and there are behavioral data for us to compare and analyze.
Ps. By the way, I suggest that the third-party statistical burying point should be a full burying point. Although we do not need to analyze all the data at the beginning, when we need analysis, we can understand and find his data without re-burying points. , Recollect data.
Phase 1: Observation period
When we confirm the modules to be modified in the next version, we must begin to prepare to observe the data template used.
Observation period: The observation period generally needs to be carried out one week to two weeks before the new version is launched. The time is too short to support the data, and the time is too long, and there is not much value. One week to two weeks is a more suitable time period.
Behavioral data is generally selected by third-party statistical systems, such as Youmeng, talkingdate, and Tencent statistics. Too much data is buried, which will increase our observation cost. Therefore, it is necessary to prepare an excel sheet template and a simple workflow.
excel template: used to extract the specified data and archive it,
Simple workflow : Export the page access and event data of the third-party statistical system into a csv file at a fixed time every day, and copy the content of the file to the excel template, and then automatically extract the data from the excel template and archive it according to the date. . Finish.
After the preparation work is completed, we only need to implement a simple workflow every day, and we can simply observe the trend of the data. This is the effect produced by the old design scheme, and the new scheme will surpass him. This is the result we all hope.
The second stage: verification period
The data needs to be verified from the first day when the new version is online. The verification here is mainly reflected in data collection. We need to collect the data generated by the new version every day.
Verification cycle: Since we have no way to guarantee the time it takes to upgrade the user version, the duration of the verification cycle will be relatively long. We divided it into two branches, numerical verification and conversion rate verification. Numerical verification requires a week of observation after the user is fully upgraded.
Conversion rate verification only requires 30% of users to upgrade before continuing to verify. If the user base is small, the upgrade rate standard can also be increased to 50% or even higher. Conversion rate verification will start early, but it will end at the same time as the value verification.
The third stage: analysis stage
Strictly speaking, the analysis phase is not carried out after the verification phase, but runs through the observation period and the verification period; just in terms of importance, it is separated into an independent module. It is emphasized again that the analysis phase is continuous, starting from the observation period and ending with the verification period.
The main task of the analysis phase is to conduct targeted data analysis. We need to interpret the meaning of the data into language and requirements, and use the data as the direction to optimize the design of the product.
The analysis stage is equivalent to data interpretation. The data dimensions provided by the third-party statistical system only include PV and UV, but they are enough for us to analyze a lot of information.
pv refers to the number of times, uv refers to the number of people, most of which apply to the number of clicks, the number of clicks, and the number of visits and the number of visitors.
Attachment: Interpretation of the meaning of some data
Overall decline: It means that the data of the new version shows a downward trend, which means that users have rebounded or lost. The reason may be that the new design style makes old users feel unfamiliar with the product and are not used to it. This problem will almost always occur in violent products. The connection between the two versions is completely cut off. In the user’s perception, this is not an optimization, but the original module is deleted and a new module is launched.
This means that users cannot find the modules they like, but they don’t like the new modules.
Overall increase: It means that the data of the new version is showing an upward trend, which means that users are very satisfied with the optimization, and a certain word-of-mouth effect has been formed, which ensures that the old users are also brought in new users. There are three main conclusions for the overall increase. One is that old users continue to use it. The other is that old users spread new users and form word of mouth. The third is that the churn rate of new users has decreased, that is, the stickiness has increased.
Regarding the third point, each module has its own churn rate, and the data of the module increases. If the increase in new users of the product is not considered, it is a manifestation of the decrease in the churn rate, that is, the retention of modules has increased.
Of course, the increase in new users of the product will also drive the increase in users of the module, and we need to eliminate some factors that affect the analysis.
Pv rise and uv fall: Pv rise refers to the increase in the frequency of users’ participation in a single day, and the fall in uv refers to the retention of the next day or the continuous decrease in stickiness. The data is more through the ratio of pv and uv, that is, pv per capita. We can induce users to produce multiple clicks in a single day through some design methods.
Of course, there are also ways to improve user retention and continued stickiness the next day.
Typical ways to improve pv:
The number of lottery draws has increased, but the rewards remain unchanged. If you can get a red envelope at one time, you will roll it up to 3 times and do it again. This design method increases user participation, but does not cause user rejection.
Typical ways to improve uv:
The number of consecutive check-ins or the cumulative check-in times for a fixed period of time, just like the Golden Box mentioned in the previous article, he gave the user a reason to use it the next day.
Of course, a certain reward mechanism is needed.
UV rises and participation rate falls: UV rises and participation rate falls. It refers to the analysis data of different objects. Most of the analysis dimensions are combined analysis of page access data and event click data.
For example, there are 100 users who visit page a, and 30 people who click the “Participate” button, and the participation rate is 30%
The participation button is the core function button of each page. There is a core button on each page. It can be the registration button for the event, or the sharing or favorite button of the article, depending on my own definition.
The core function button is the most desired operation button for a certain page.
The appearance of this kind of data indicates that the conversion ability of the core button has decreased, and the reasons are as follows:
1. The design is not prominent enough, and the user failed to find this button in the first time
2. The design is ambiguous, and the user failed to understand the button in the first time
The funnel effect is a commonly used technique. It is not only applied to the path, such as building a funnel based on the user’s visual stay time.
like:
The user stays on page A for 4 seconds
Is the core button discovered by the user in the first second or by the user in the fourth second?
The conversion rate of the former is often dozens of times higher than the latter.
By the way, the registration button on the event page will be placed in the lower middle of the first batch, so that users can find him in the second second.
Because in the first second, users have to see the theme of the event, so that they can form a motivation to participate.
